gedlee said:
Hi Earl,
The intend it so show the fundamental differences in the various woofer systems. Room effects and room placement, listening positions are very complex, as you know, for any woofer. We seldom have even a closed room. There are always openings to adjacent room, door, hallways, particularly in newer homes with open floor plans. Every istallation is different.
Placement and interaction with room modes is of course the issue though. Dipoles have some unique characteristics due to their inability to excite the room significantly below the room fundamental. As a result they sort of "tune' themselves to the room. But, referring to Backman (2003), dipoles have the most serious placement problems. My experimentation with the CRAW woofer system seemed to agree. With proper placement they can excite fewer room modes above the fundamental, but this isn't necessarily a good thing as those fewer excited modes become more dominant in my experience. Still, many people love dipole bass. Cardioids seem to be least sensitive to placement and listening position, according to Backman, and again my experiments seem to agree. However, below the room fundamental monopoles are superior if you are trying to take advantage of room pressurization effects. The biggest problem with monopole woofers, IMO is that they are often poorly matched with the room. Everyone wants deep powerful bass so they buy/build woofers with Fs well below the room fundamental. That coupled with room pressurization effects is a recipe for boomy bass since the woofer will over power the room (maybe desirable for HT, but not music, IMO). I try to build sealed woofers so that Qts = 0.5 with Fs at or slightly above the room fundamental. That way both the woofer and room compliment each other.
With the systems I offer, the NaO II and the NaO Mini I have taken two different approaches. With the NaO II the woofer system is a quasi-cardioid (acoustic resistance box) but it can be sealed at the back to form a sealed box when high low frequency SPL is required. Equalization for the cardioid and sealed box formats are controlled electronically in the active woofer/panel crossover. In cardioid mode I have been very pleased with the behavior of the system when auditioned in a variety of different rooms.
In the NaO Mini I took a different approach. The woofers are sealed boxes but I use an active circuit for pole shifting the woofer system so that the system fs can be set based on the installation. With correct woofer placement the woofer/ panel crossover results is a smooth transition from dipole in the mid frequencies to cardioid through the crossover region to monopole at low frequency. I like the idea of the Mini because it keeps the bass very clean sounding while retaining good max SPL capability without the need for the equalization required for a gradient woofer, and keeps cost down. Additionally, although it negates the cardioid nature of the response through the pane/woofer crossover, the separate woofer enclosures allow the woofer to be moved around independently of the main panel adding additional flexibility for difficult installations. The idea with the Mini is to place the panels for optimum midrange presentation and then place the woofer according to the design criteria. This should yield the best presentation form the upper bass region on up. But if necessary, they woofer placement can then be altered to the listener’s preferences for deep bass response.
john k... said:
Do you have a reference for Backman (I assume Juha).
I am always disturbed by statements like "excite fewer room modes" which is virtually never the case and wouldn't be a good thing if it were. What we want is smooth non-resonant (boomy) bass, and NOT exciting modes is a sure way to make it boomy. The modes are unavoidable. You may be able to not excite one, but then you are going to excite another and in the end it all comes down to the same thing. Exciting as many modes as possible will always yield the smoothest response, no matter what kind of source you use. It is, of course, a good idea to have those modes well damped, but this is universally true regardless of source type, location, listener, etc. Using more LF sources excites more modes, yields a smoother response and is almost universally agreed to as the optimal answer. Various authors may disagree on the details of "optimal number", "optimal locations", etc., but we all agree that more is better. The type of source will only affect the response for the bottom several modes and below down to DC, and, as you correctly point out, monopoles will always win here. Thats what guides my choice.
Hence, it seems to me that talking about how any ONE source may work in any given room, is not really the point. It certainly doesn't matter if multiple sources are used dince placement, source type, etc. all become blurred as the number of sources goes up - they all lead to the same result.
I use bandpass monpoles for higher output and lowered distortion. I try and tune them all to different frequencies away from the room modes. The sources resonant frequency is another "pole" (mathematically - a resonance) and its a good idea to distribute these resonances as uniformly as possible, especially at very LF where there are so few of them. A high Q source tuned at the same frequency as a high Q room mode will be a sure disaster. I agree with your lower Q approach as, in general, I do this too. Peaked bass, as so often used in stores for "buyer impact" is often a disaster when you get them home.
I just don't see all this discussion of types as being the main point. I see numbers as the important issue and it seems to me that the type enters the problem only in regards to the number of actual drivers that one has to buy. For example, I could certainly live with three dipole subs as long as there were some monopoles to handle the very lowest frequencies that the dipoles cannot drive. But this seems to me to simply be using more drivers than necessary. Its not the most cost effective approach for comparable bass smoothness. The cardiod is one step better than the dipole - if it works as advertised, but it still needs the monopole down low.
If you are talking about the response from a single source at LF then you could be correct, although my published results don't quite agree, but to me that discussion is not the important one since I would never use a single sub in any room.
One more thing.
The whole idea of directivity at very low frequencies is incorrect. Sound waves, at a mode, cannot freely travel in space in any direction. They must travel through the room in very percise unalterable directions no matter what the sources directivity may be. If multiple modes are excited then multiple directions are possible. In any space where there are not a sufficient number of modes to allow for a "free space" (arbitrary direction) type of propagtion - such as the LF in any home size room - the concept of directivity is completely inapplicable. This is exactly what I mean when I say that the source and the room couple. The source cannot behave in a room the same way that it would in free space. The room wil completely dominate what it ALLOWS the source to do. So all of these nice polar diagrams for directivity ONLY apply outside of the room and have no relavence to how that source will work in a real room.
The whole idea of directivity at very low frequencies is incorrect. Sound waves, at a mode, cannot freely travel in space in any direction. They must travel through the room in very percise unalterable directions no matter what the sources directivity may be. If multiple modes are excited then multiple directions are possible. In any space where there are not a sufficient number of modes to allow for a "free space" (arbitrary direction) type of propagtion - such as the LF in any home size room - the concept of directivity is completely inapplicable. This is exactly what I mean when I say that the source and the room couple. The source cannot behave in a room the same way that it would in free space. The room wil completely dominate what it ALLOWS the source to do. So all of these nice polar diagrams for directivity ONLY apply outside of the room and have no relavence to how that source will work in a real room.
Earl,
James Kates wrote a paper entitled "Dipole Loudspeaker Response in Listening Rooms" that shows improved in-room low-frequency smoothness for a dipole vs a monopole.
As I seek to reconcile his findings with your work and with Welti's, this thought comes to mind: A dipole can be thought of as two monopoles, one with reversed polarity, separated by a path length - thus one would expect increased bass smoothness based on the principle of smoothness increasing as the number of in-room low frequency sources goes up. However, that might only apply to independent low frequency sources; a conventional reflex system can also be thought of as two (non-independent) low frequency sources at frequencies where the woofer and port are both operational.
Duke
James Kates wrote a paper entitled "Dipole Loudspeaker Response in Listening Rooms" that shows improved in-room low-frequency smoothness for a dipole vs a monopole.
As I seek to reconcile his findings with your work and with Welti's, this thought comes to mind: A dipole can be thought of as two monopoles, one with reversed polarity, separated by a path length - thus one would expect increased bass smoothness based on the principle of smoothness increasing as the number of in-room low frequency sources goes up. However, that might only apply to independent low frequency sources; a conventional reflex system can also be thought of as two (non-independent) low frequency sources at frequencies where the woofer and port are both operational.
Duke
gedlee said:
Well I think a number of researchers have established that, depending on placement and orientation, a dipole can be shown to excite fewer modes, but I agree that this is not necessarily a good thing. That is what I said previously.
I am basically in agreement here but I believe that the Backman work shows that in the modal region the cardioid tends to excite the various modes more uniformly. That seems to be the case with my (limited) measurements and also in some of my own simulations. Above the modal region there isn't much difference between sources, other than how they control the first reflection, which does has a significant impact on the perceived sound field, as I am sure you are aware. Below the modal region the monopole is supreme, agreed, provided it is correctly coupled to the room, but the cardioid still has some of the monopole characteristics of room pressurization as monopole as DC is approached, though to a lesser extent.
Well here I would disagree, but primarily because I don't think the average listener has the option to employ a distributed array of low frequency sources. Practically one or two sources seems more reasonable. I would argue that stereo sources are the minimum acceptable and likely the format of the majority of how environments. Thus, how one or two sources works in a given room is a very real concern.
In any event, the Backman reference is "Low frequency polar pattern control for improved in-room response", AES convention paper 5867, 115th AES convention, Oct, 2003.
john k... said:
You'd have to give me a reference for that, because I don't believe that one exists, especially because its not true. I'm talking about showing this scientifically (math works for me), not some hand-waving hueristic.
A dipole will not excite fewer modes than a monopole at the same location. It excites them differently, but not fewer of them. And if they both excite the same number, but differently, then one can only reasonably talk about differences in the excitation averaged across some bandwidth - which is exactly what I do in my study. And this shows that a dipole DOES NOT have a lower variation in excitation, only a lower total excitation.
One cannot pick a spot where the dipole excites fewer modes or at a lower level than a monopole, ignoring those spots where the opposite is true, and call it a proof. You must give both sources the same chance.
audiokinesis said:
To be considered as "statistically" independent sources they have to be "well" seperated. Two sources close together are completely correlated (if they are fed the same signal) and not independent - they only count as one. Thus a dipole looked at in this way is only one source not two. As the distance between the sources gets greater the correlation drops and they become more independent. Thats why I recommend placing them as far apart as possible.
This displaced opposite polarity point sources is exactly the way that a dipole is modeled in the analysis that I do.
Hi Earl,
How do you propose to use multiple monopole sources throughout the room below 200-300hz - with stereo material, unlocalized.
Also, if you could be so kind as to answer this question that you may have missed in that long thread. Thanks.
cheers,
AJ
How do you propose to use multiple monopole sources throughout the room below 200-300hz - with stereo material, unlocalized.
Also, if you could be so kind as to answer this question that you may have missed in that long thread. Thanks.
cheers,
AJ
I would like to rephrase AJinFLAs question, since i thought about the same problem...
Up to which frequency do you use your subwoofer? Looking at the Summa, it seems that a crossover around 50-80hz seems to be the case, but this is only guessing.
In my room, i feel that the modes below around 80hz are the worst, so getting rid of them would be really great. But what about the modes above this? There seem to be modes even higher than 200hz. I dont want to compromise on the stereo image, so multiple subs are no solution here. How do you work around this problem?
Up to which frequency do you use your subwoofer? Looking at the Summa, it seems that a crossover around 50-80hz seems to be the case, but this is only guessing.
In my room, i feel that the modes below around 80hz are the worst, so getting rid of them would be really great. But what about the modes above this? There seem to be modes even higher than 200hz. I dont want to compromise on the stereo image, so multiple subs are no solution here. How do you work around this problem?
AJinFLA said:
MaVo said:
First to the LF question:
I don't use subs up as high as 200 Hz. ever. The Summas and most of my designs go down to 50 - 70 Hz or so, hence a sub is necessary to fill in below the mains cutoff. I usually make this a narrow bandpass which yields a lot of output in a small bandwidth and small package. At Ai (www.ai-audio.com) we called these ULF subs. Then I distribute two more Broad Band (BB) subs around the room which are also bandpass, but tuned to have a broad bandwidth such that they overlap the ULF and the mains. These go up to about 150 Hz or so. Of course there is always some fitting of the subs into the specific room thats why I can't give you the exact numbers. (Someday I'll write-up how to correctly setup multiple subs in a room.) The BB subs are usually so low in level that you can't even tell there on unless you are right on top of them, so localizing on them is never an issue.
Think of it this way. The mains and the ULF carry the bulk of the energy, but there will be the inevitable holes in the response. Remember (or maybe you didn't know this) that holes are much deeper than peaks are high, its almost 2:1 (+6db to -11 dB on average). So to get good bass you have to fill in these holes or the bass sounds empty. Raising the bass level only makes it sound boomy. So in essence the BB subs ONLY fill in the holes and this is not really a lot of energy, but it makes a huge difference in both the measured response and the sound quality.
As to the Klippel test. These examples are set up to be obvious for obviuos reasons. The nature of the distortion doesn't change so when you "dial up" the level there will be a 1 to 1 correlation between the audiblity and the level. No big secret there. I could just as easily show examples of different forms of distortion where the numbers had absolutely no relavence to the audibility(and I did post these on my web site sometime back) . The .1% THD sounded awful and the 15% was inaudible. Its all in how you set up the examples and the tests. (I got disgusted with that other thread so I left.)
To find truth you have to look for the truth. Its always easy to setup a test to prove anything that you want to prove - you just ignore all of the contradicting evidence and examples. It happens all the time.
gedlee said:
Proof is only a mathematical concept. Noting needs to be proved here. This is an optimization problem. One doesn't say to a listener "you must place every speaker you listen to at the same position to make a fair comparison". Quite the contrary, any resonable listener will position each speaker to optimize the response. The question is then, what type of woofer system yields the best response in the modal region with optimum placement? The premise is that with optimal placement smoother response can be obatined with a single dipole source than a single monopole source. There is no restriction that the placement must be the same. This seems to be the conclusion of Kates and and also of Jorma Salmi (AES preprint #9927, 1992). Backman seems to be saying that the cardioid is superior to either since it is less sensitive to placement. My experiences are certainly in agreement. And I think you will find tha most people who have systems with either cardioid or dipole bass will tell you the same thing. The optimization problem has a better solution in the modal region when the source is other then monopole. And seems to have the best soultion with a cardioid.
But there are other issues as well, total radiated power being one. The difference between the behavior of a perssure source and a velocity source being another (with the cadrioid somewhere in between). Some of this is easily understood by considering the placement of the source on a ground plane and then considering the effect of the distance between the source and a reflecting wall behind the source. When the distance is less than a wave length the monopole response is augmented by the image source, ultimately increasing output in the front hemisphere by 6dB, for the dipole the image source cancels and bass response is decreased and ultimately nulled. For the cardioid the image source contributes nothing to the on axis response and the response degenerates to that of a monopole inthe front hemisphere but with no change in the on axis level. Of course, in a real room the results will be somewhat different because the typical wall behind the speaker is not perfectly reflective at low frequency and the image source would effectively be attenuated relative to the real source.
Of course, these are aguments relating to optimal set up of a system for a single listener. Other considerations would be introduced if the discussion were directed at a wide range of listening positions which might be the case for a HT setup.
john k... said:
This completely misses my point. I am saying that one must compare the sources OPTIMALLY setup, not one setup optimally and the other sub-optimally.
I completely concur with the question, its was exactly what I studied in my AES paper (whcih was reviewed and published, unlike the preprints that you are quoting) This WAS NOT the conclusion in my paper - why don't you quote that? You can't just pick and choose your data sources. Thats not professional.
If you are talking about using a single sub this is not as clear and I'd say the monople would still win, but with multiple subs, as recommended, the monopole would clearly be the optimal because of its higher efficiency at low frequencies.
Your example here is incomplete because "in a real room the results will be" completely "different", which is why I used a real room in my study. You can't just use these one wall, corner, etc analogies at these LFs, they don't work. There are six walls and you need to consider all six of them.
John - I realize that you sell cardiod subs and that you have a strong reason to prefer them, but it is not right to simply throw out data that does not support your position. And admittedly I will have to read Juha's paper as he generally does good work to see what he says. But remember that I continue to say that using multiple subs makes this discussion academic.
gedlee said:
I don't have your paper but would be glad to consider it. However, let's not harp on the "reviewed and published" bit. I've been around long enough and presented/published/reviewed enough technical papers to know what the process is and its significance. I've seen more technically inadequate papers reviewed and published in respectable journals in my career than I would care to admit, some of then just outright wrong from the get go. Like everything else, it's largely a political environment.
Actually it would seem more likely that the discussion of multiple subs is the academic one since it is far more unlikely a scenario than a typical set up with a single or stereo subs. It may be the best solution, but highly impractical and far more difficult to set up optimally. For a give room, how many and where, not to mention all the additionally (cost) amplification, crossovers...... required, and didn't you mentioned de-correlators? Developing an algorithm for placement and number of sources to produce optimal response in any room is an interesting problem, academically, but it just doesn't offer a practical solution in the majority of cases.
john k... said:
Please remember that the pro installation industry product volume and service or renewal rate is such, that dwarfs today's Hi-Fi industry size. There, and in true HT, whatever works is adopted. 2 distributed WB subs and a narrow band bigger one plus a couple of dsp cross channels are peanuts as an economic or technical requirement in clubs and dedicated HT rooms.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.